In nuclear fusion reactors, tritium dynamics plays a dominant role. An unprecedented amount of tritium is consumed in Deuterium-Tritium (D-T) nuclear fusion reactors, ~0.5 kg per day for 3 GW fusion power. However, tritium is radioactive, has short half-life (~12.33 years), and is present in nature in negligible concentration. Because of tritium scarcity, future fusion power reactors must be self-sufficient, i.e. the reactor must have a closed fuel cycle where tritium is produced in greater amounts than it is consumed. Furthermore, nuclear fusion reactors must accumulate and provide tritium start-up inventory for the next generation of fusion power plants, since natural reserves of tritium are very limited. Moreover, because of its radioactive nature, tritium presents a serious hazard to the personnel and has implications to safety and nuclear licensing. Accurate predictive models of the nuclear fusion fuel cycle are required to effectively design the fuel cycle components, understand tritium dynamics in the fusion fuel cycle, and determine the technology and physics requirements to attain tritium self-sufficiency. Moreover, accurate predictions of tritium inventories and flow rates within fusion components, and estimations of tritium releases to the environment are necessary for nuclear licensing. In this dissertation, two numerical models are developed to perform tritium transport assessment within fusion systems. First, a high fidelity numerical model is developed to simulate time-dependent tritium transport within the reactor outer fuel cycle (OFC). Detailed (high resolution) component-level models, where constitutive transport equations are implemented in COMSOL Multiphysics and solved for various fusion sub-systems, are integrated into system-level with the use of MATLAB/Simulink S-Functions to reproduce typical OFC tritium streams. The model is applied to the KOrean Helium Cooled Ceramic Reflector Test Blanket System (KO-HCCR TBS) which will be tested in the International Thermonuclear Experimental Reactor (ITER). However, the developed model offers some flexibility and can be applied to other Test Blanket Module (TBM) designs. Second, the overall fusion fuel cycle is modeled analytically by a system of time-dependent zero-dimensional ordinary differential equations with the tritium mean residence time method. This technique yields results useful for understanding the overall fuel cycle dynamics and the importance of certain components and parameters. The analysis of tritium inventories and flow rates is extended to determine the physics and technology requirements to attain tritium self-sufficiency. In particular, the state-of-the-art plasma physics and technology parameters (e.g. tritium burn fraction, fueling efficiency, processing times, etc.) and up-to-date fuel cycle design are considered in the analysis. The tritium self-sufficiency assessment and tritium start-up inventory evaluation are performed to investigate: (i) the effect of the reactor operating scenario and availability factor, e.g. to account for random failures and ordinary maintenance, (ii) the scenarios for commercialization, e.g. risk associated with tritium reserve inventory reduction, (iii) the penetration of fusion energy into power market, e.g. effect of the doubling time, and (iv) the effect of reactor power on tritium start-up inventory, e.g. for plasma-based test facilities, DEMOnstration reactors (DEMO), and power reactors. The results highlight the physics and technology R&D requirements to attain fuel self-sufficiency in fusion reactors.